Oligofructose alone and in combination with 2′fucosyllactose induces physiologically relevant changes in γ-aminobutyric acid and organic acid production compared to sole 2′fucosyllactose supplementation: an in vitro study

Abstract We explored the potential for the prebiotic oligofructose and prebiotic candidate 2′fucosyllactose, alone and in combination (50:50 blend) to induce physiologically relevant increases in neurotransmitter (γ-aminobutyric acid, serotonin, tryptophan, and dopamine) and organic acid (acetate, propionate, butyrate, lactate, and succinate) production as well as microbiome changes using anaerobic pH-controlled in vitro batch culture fermentations over 48 h. Changes in organic acid and neurotransmitter production were assessed by gas chromatography and liquid chromatography and, bacterial enumeration using fluorescence in situ hybridization, respectively. Both oligofructose and oligofructose/2′fucosyllactose combination fermentations induced physiologically relevant concentrations of γ-aminobutyric acid, acetate, propionate, butyrate, and succinate at completion (all P ≤ .05). A high degree of heterogeneity was seen amongst donors in both neurotransmitter and organic acid production in sole 2′FL fermentations suggesting a large responder/nonresponder status exists. Large increases in Bifidobacterium, Lactobacillus, and Bacteroides numbers were detected in oligofructose fermentation, smallest increases being detected in 2′fucosyllactose fermentation. Bacterial numbers in the combined oligofructose/2′fucosyllactose fermentation were closer to that of sole oligofructose. Our results indicate that oligofructose and oligofructose/2′fucosyllactose in combination have the potential to induce physiologically relevant increases in γ-aminobutyric and organic acid production along with offsetting the heterogenicity seen in response to sole 2′fucosyllactose supplementation.


Introduction
In recent years, there has been significant growth in interest in the bidir ectional r elationship existing between the gut and the brain-a term coined the gut-brain axis (Appleton 2018 ).There is an increasing body of evidence suggesting that the composition of the gut microbiota plays a k e y role in influencing mood state including emotional r egulation, cognitiv e performance, and mental health, particularly anxiety and depression (Evrensel and Ceylan 2015 ).Within the gut, it is documented that se v er al gener a, species and strains of bacteria can produce a number of different metabolites associated with cognitive and mental state including neurotr ansmitters suc h as γ -aminobutyric acid (GABA), ser otonin, and dopamine as well as organic acids-acetate , propionate , butyrate , lactate, and succinate (Cryan et al. 2020, Silva et al. 2020 ).
Pr e v alent GABA pr oducers in the gut include se v er al species and strains of Bifidobacterium and Lactobacillus and to a lesser extent Bacteroides (Str andwitz 2018, Str andwitz et al. 2019 ) .Studies hav e r eported that GABA can be pr oduced via se v er al pathways: first, by the conversion of arginine to ornithine to putrescine and GABA, resulting in the generation of succinate and finall y pr opi-onate (Otaru et al. 2021 ).Second, it can be produced by the decarboxylation of glutamate (Barrett et al. 2012 ).GABA serves as an acid-r esistance mec hanism to surviv e the low pH of the intestinal envir onment (Str andwitz et al. 2019 ).Ho w e v er, it r emains unclear whether GABA is able to cross the blood-brain barrier (Knudsen et al. 1988, Shyamaladevi et al. 2002 ), although it can be synthesized from acetate in the hypothalamus (Frost et al. 2014 ).Incidentall y, ther e is accumulating evidence demonstrating that bacterial derived GABA more likely functions locally, being involved in neuronal excitability within the enteric nervous system along with contributing to w ar ds GI motility in the ilium and peristaltic reflux within the colon (Seifi et al. 2014 , Seifi andSwinny 2019 ).
Another neur otr ansmitter, ser otonin, is pr oduced via the metabolism of tryptophan in the presence of enter oc hr omaffin cells by various microorganisms including Lactococcus , Enterococcus , and Streptococcus (Kaur et al. 2019 ) .Yet, it is estimated that ∼90%-95% of tryptophan enters the kynurenine pathway resulting in the generation of either kynurenic or quinolinic acid (Gao et al. 2020, Muneer 2020 ).Of these tw o metabolites, k ynurenic acid has been associated with neur opr otectiv e pr operties and r egula-tion of immune function, whereas increasing concentrations of quinolinic acid in the brain is associated with psychiatric and neur odegener ativ e disorders (Lugo-Huitron et al. 2013 ).
Or ganic acids suc h as acetate and lactate are predominately produced by Bifidobacterium via saccharolytic fermentation are reported to act as endocrine signalling molecules (Silva et al. 2020 ).Butyr ate pr oduced by bacteria including Bacteroides and Roseburia , either dir ectl y or as a r esult of cr oss-feeding on acetate and lactate, is speculated to aid in the expression of GABA receptors (Nank ov a et al. 2014 ).Furthermore, as documented by both (Reigstad et al. 2015, Dalile et al. 2019 ) organic acids likely aid in regulating the expression of tryptophan-5-hydroxylase 1 and tyr osine hydr oxylase , enzymes in volv ed in the r ate-limiting step in the biosynthesis of se v er al neur otr ansmitters by enter oendocrine cells.
The majority of glutamate and tryptophan, like many other amino acids, are absorbed and metabolized in the small intestine.Yet, a r eport fr om (Yao et al. 2016 ) estimated that on av er a ge between 7% and 10%, or 6-18 g/day, of dietary pr otein r eac hes the large intestine intact.This roughly equates to 3.4-6.3mmol/kg total free amino acids entering the proximal and distal region of the colon.Curr entl y, glutamate is estimated to be one of most abundant amino acids present in food, making up between 8% and 10% of all dietary protein consumed (van der Wielen et al. 2017 ).Giv en the shar p rise in high quality protein diets and glutamate ric h foods (Beyr euther et al. 2007, Tennant 2018 ) it is estimated that daily consumption of glutamate ranges from 5 to 15 (12 aver a ge) g/day.In Asian countries this is lar gel y due to increased consumption of free monosodium glutamate.Based on the assumption that 7%-10% of total dietary pr otein r eac hes the colon intact, and an av er a ge glutamate intake of 12 g/day, it can be speculated that some wher e in the region of 0.7-1.2g of total dietary glutamate r eac hes the colon.Dail y tryptophan consumption is estimated at ∼900-1000 (950 av er a ge) mg/day (Ric hard et al. 2009 ) r oughl y equating to between 66.5 and 95 mg dietary tryptophan entering the colon.
The majority of r esearc h inv estigating c hanges in neur otr ansmitter production via manipulation of the microbiota primarily r e volv es ar ound se v er al str ains of pr obiotics .T his makes sense giv en the r emarkable v ariability in neur otr ansmitter pr oduction seen in micr oor ganisms, e v en those within the same genera and species (Kaur et al. 2019, Strandwitz et al. 2019 ).Ho w e v er, other means of targeted manipulation of the gut microbiota exist, including prebiotics and potential prebiotic oligosaccharide candidates .T he most substantiated of all prebiotics are oligofructose (OF) and inulin, which belong to a class of nondigestible carbohydr ates r eferr ed to as inulin-type fructans (ITF) (Mensink et al. 2015 ).To date, the ability of ITF to stim ulate c hanges in gut microbiota composition has been substantially demonstrated, both in vivo and in vitro , across a wide array of dosages (Wang and Gibson 1993, Kolida et al. 2007, Vandeputte et al. 2017 ).Other oligosaccharides under consideration as prebiotics include human milk oligosaccharides (HMOs), a group of structurally diverse and complex unconjugated glycans present in human br east milk (Ninonue vo et al. 2006 ).The most common of these is 2 fuscosyllactose (2 FL), the first HMO to be produced on an industrial scale (Sprenger et al. 2017 ) and curr entl y under inv estigation as a novel food ingredient and as a means of treating IBS and cognitive mental disorders (Al-Khafaji et al. 2020, Sans Morales et al. 2022 ).In comparison to ITF, the data on the ability of 2 FL to stim ulate c hanges in micr obial composition ar e mixed due to a limited number of clinical studies (Elison et al. 2016, Suligoj et al. 2020, Iribarren et al. 2021, Ryan et al. 2021 ).
Muc h r emains unknown on whether supplementation with pr ebiotics/pr ebiotic candidates is enough to stimulate physiologicall y r ele v ant incr eases in neur otr ansmitter pr oduction.This is in part due to the large heterogeneity existing between individual gut microbiotas and the need for an individual's microbiota to possesses the r equir ed micr oor ganisms.As a r esult, in this study we investigated whether it is possible for the prebiotic OF and prebiotic candidate 2 FL, singular and in combination, to stimulate physiologicall y r ele v ant incr eases in neur otr ansmitter and or ganic acid pr oduction using pH-contr olled in vitro batc h cultur e fermentation over 48 h.

Reagents
Unless otherwise stated all r ea gents used in this experiment were sourced from Merck, Gillingham, UK.

Faecal sample preparation
Ethical a ppr ov al of collecting faecal samples from healthy volunteers was obtained from University of Reading Research Ethics Committee.Fr eshl y voided faecal samples were obtained from five healthy adults aged between 18 and 40, who had not taken antibiotics for at least 4 months prior to the experiment, had no history of gastrointestinal disor ders, w ere not regular consumers of prebiotics or probiotics and did not follow any restrictive diet.Faecal samples were diluted 1 in 10 (w/v) using 0.1 mol/l, pH 7.4 anaerobicall y pr epar ed phosphate buffer ed saline (PBS, Oxoid, Hampshir e, UK).Faecal samples were then homogenized in a stomacher (Sew ar d, stomacher 80, Worthing, UK) for 120 s at 260 paddle-beats per min.Faecal slurry (5 ml) was immediately used to inoculate eac h batc h cultur e v essel.

pH-contr olled, stirr ed batch culture fermentation
For each donor one independent batch culture was run.For each, batc h cultur e v essels (4 × 100 ml) were set up and 45 ml of basal nutrient media wer e asepticall y pour ed into eac h v essel.This system was left overnight with oxygen free nitrogen pumping through the medium at a rate of 15 ml/min with constant agita-tion and this continued throughout the course of fermentation.Before adding the faecal slurry, a water bath was used to set the temper atur e of the basal medium at 37 • C, and a pH of between 5.4 and 5.6 was maintained throughout the course of fermentation using a pH meter (Electrolab pH controller , T ewksbury, UK) via the addition of 0.5 mol/l HCl or 0.5 mol/l NaOH.Stirring of faecal samples was maintained using a ma gnetic stirr er.For eac h donor thr ee differ ent substr ates wer e pr epar ed with one v essel containing each of the following substrates (1% w/v): OF, 2 FL, and OF + 2 FL (50/50 w/w).One vessel was set up as the negativ e contr ol with no added carbohydr ate.All v essels wer e inoculated with 5 ml of a 10% (w/v) faecal slurry (diluted with PBS).A sample (4 ml) was r emov ed fr om eac h v essel after 0 h, 10 h , 24 h, and 48 h incubation to ensure enough sample was taken for bacterial, organic acid, and neur otr ansmitter anal ysis by fluor escence in situ hybridization-flo w c ytometry (FISH-FLOW), gas c hr omatogr a phyflame ionization detection (GC-FID), and triple quadruple liquid c hr omatogr a phy-mass spectr ometry (LC-MS QQQ), r espectiv el y.

Enumer a tion of faecal microbial populations by flow cytometry fluorescence in situ hybridization (FISH-FLOW)
A 750 μl sample of batch culture fermentation effluent was centrifuged at 11 337 × g for 5 min.The supernatant was then discarded, and the pellet was then suspended in 375 μl filtered 0.1 mol/l, pH 7.4 PBS solution.Filtered 4% paraformaldehyde (PFA) at 4 • C (1125 μl) were added and samples were stored at 4 • C for 4 h.Samples were then washed thor oughl y with PBS three times to remove PFA and resuspended in 150 μl PBS and 150 μl 99% ethanol.Samples were then stored at −20 • C, until FISH analysis by flow cytometry could be conducted.The probes used in this study are presented in Table 1 .
A volume of 75 μl of fixed samples were mixed with 500 μl filtered cold (4 • C) 0.1 mol/l, pH 7.4 PBS and then centrifuged at 11 337 × g for 3 min.The resulting supernatant was then discarded, and pellets resuspended in 100 μl of TE-FISH (Tris/HCl 1 mol/l pH 8, EDTA 0.5 M pH 8, and filtered distilled water with the ratio of 1:1:8) containing lysozyme solution (1 mg/ml of 50 000 U/mg protein).Samples were then incubated for 10 min in the dark at room temperature and centrifuged at 11 337 × g for 3 min.Supernatants w ere discar ded, and pellets w ashed with 500 μl filtered cold PBS by aspiration to disperse the pellet.Samples were then centrifuged at 11 337 × g for 3 min and supernatants discarded.
P ellets w er e r esuspended in 150 μl of hybridization buffer, aspirated using a pipette and gently vortexed.Samples were centrifuged at 11 337 × g for 3 min and supernatants discar ded.P ellets were resuspended in 1 ml of hybridization buffer.Aliquots (50 μl) of samples were placed in labelled 1.5 ml Eppendorf tubes and 4 μl of specific probes (50 ng/ μl) were added.Samples were incubated at 35 • C for at least 10 h in the dark.
Following incubation, 125 μl of hybridization buffer were added to each tube and gently v ortexed.Samples w ere then centrifuged at 11 337 × g for 3 min and supernatants were discarded.Pellets w ere then w ashed with 175 μl of washing buffer solution and gently vortexed.Samples were incubated at 37 • C for 20 min and centrifuged at 11 337 × g for 3 min.Supernatants w ere discar ded and different volumes of filtered cold PBS (300, 600, and 1200 μl) were added based on flow cytometry load.Samples were k e pt at 4 • C in the dark until flow cytometry measurements could be conducted.Fluor escence measur es wer e performed on an BD Accuri™ C6 Plus (BD, Erembodegem, Brussels) measuring at 488 and 640 nm.A threshold of 9000 in forw ar d scatter (FSC-A) and 3000 in side scatter (SSC-A) was placed to discard bac kgr ound noise, a gated area was applied in the main density dot to include 90% of the e v ents.Flow rate was 35 ul/min, limit of collection was set for 100 000 e v ents and anal ysed with Accuri Cflo w Sampler softw are.Bacterial counts were then calculated through consideration of flow cytometry reading and PBS dilution.

Organic acids by gas chromatography-flame ionization detection (GC-FID)
Samples (1.5 ml) of batc h cultur e fluid were collected and centrifuged at 11 337 × g for 10 min.Supernatants were transferred to 1.5 ml Eppendorf tubes and stored at −80 • C until analysis could be conducted.Sample extractions were performed according to (Richardson et al. 1989 ) with modifications.Briefly, 1 ml of sample was tr ansferr ed into a labelled 100 mm ×16 mm glass tube (International Scientific Supplies Ltd, Bradford, UK) and 50 μl of 2-ethylbutyric acid (0.1 mol/l, internal standard), 500 μl concentrated HCl and 3 ml diethyl ether were added to each glass tube before vortexing for 1 min.Samples were centrifuged at 2000 × g for 10 min.The resulting diethyl ether (upper) layer of each sample was tr ansferr ed to clean 100 ml scr e w top glass tubes.Ether extract (400 μl) and 50 μl N -tert-butyldimethylsilyl)-N -methyltrifluor oacetamide (MTBSTFA) wer e added into a GC scr e w-ca p vial.Samples were left at room temperature for 72 h to allow samples to completely derivatise.
An Agilent/HP 6890 Gas Chr omatogr a ph (He wlett P ac kard, UK) using an HP-5MS 30 m × 0.25 mm column with a 0.25 μm coating (cr oss-linked (5%-phen yl)-methylpol ysiloxane, He wlett P ac kard) was used for analysis of or ganic acids.Temper atur es of injector and detector were 275 • C, with the column temperature progr ammed fr om 63 • C to 190 • C at 15 • C/min follo w ed b y 190 • C for 3/min.Helium was the carrier gas (flow rate 1.7 ml/min; head pr essur e 133 KP a).A split r atio of 100:1 was used.Quantification of organic acids was achieved by calibration with acetic , propionic , butyric , lactate , and succinate acids in concentrations between 12.5 and 100 mmol/l.Mean metabolite concentrations were expressed as mmol/l.

Neur otransmitter pr oduction by LC-MS QQQ
Samples (500 ul) of batch culture effluents were diluted to 1/100 and 1/1000 in LC-MS grade water and 1 ml was pipetted into HPLC scr e w top vials.For quantification of neur otr ansmitters a Shimadzu QQQ equipped with a Discovery HS F5 HPLC Column (3 mm particle size, L × I.D. 15 cm × 2.1 mm) maintained at 40 • C was used.The mobile phase comprised of solvent A (0.1% v/v formic acid) and solvent B (acetonitrile containing 0.1% formic acid) at a flow rate of 0.25 ml/min.The gradient elution program w as as follo ws: 2-5 min solv ent B fr om 0% to 25%, 5-6 min solvent B from 25% to 95%, then holding for 2 min, 8-9 min from 95% to 0% and then until 15 min.
A LC/MS-8050 triple quadrupole (QQQ) detector was operated in the multiple reaction monitoring (MRM) mode using the polarity-switc hing electr ospr ay ionization mode.Dry gas temperature was set at 200  ( 1996 ) ac hie v ed via generation of a linear calibration curve ranging from 1.00 to 1000 ng/ml based on the detected signal proportional to concentration of the analyte.Good linearity of fit was considered as an R 2 of greater than 0.99.

Sta tistical anal ysis
Statistical P ac ka ge for Social Science v ersion 27 (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses.Changes in bacteriology, neur otr ansmitter and or ganic acid pr oduction wer e analysed using a general linear model (GLM) to assess repeat measures.Post hoc comparisons were also performed in order to determine any significant differences between interventions at 48 h in organic acid and neurotransmitter production.All post hoc pairwise comparisons were corrected for type 1 errors using Bonferroni adjustment within each GLM.All tests w ere tw o tailed and P -v alues wer e consider ed significant at P ≤ .05 and are displayed by specified P -v alues.Gr a phs wer e gener ated in Gr a phP ad Prism version 10.0.0 for Windows, GraphPad Software, San Diego, California, www.graphpad.com .

Neur otransmitter pr oduction
Figur e 1 r eport the concentr ations of GABA, ser otonin, tryptophan, and dopamine over the course of the 48 h fermentation.Nore pine phrine, e pine phrine, and kyn ur enic acid wer e below the limit of detection and were therefore excluded from analysis.Mean and specific donor data ar e pr esented in Tables S1 and S2 ( Supporting Information ).
The lar gest c hanges seen in neur otr ansmitter pr oduction acr oss all substr ates tested wer e r ecorded for GABA.The extent of change and fermentation patterns varied substantially across the substrates tested.Of all substrates tested, OF induced the largest increases in GABA at 48 h going from 3605.01 ± 1347.35 standard error (SE) ng/ml to 836 187.28 ± 303 310 (SE) ng/ml (832 582.27 mean difference; P = .004).In contrast, increases in GABA concentration in the 2 FL fermentations were highly heterogenous amongst donors, subsequently inducing the smallest av er a ge increase at 25,5763.59 ng/ml ± 84 953 (SE) above baseline at 24 h (Fig. 1 A; Table S1 , Supporting Information ).
Figure 1 (B) and (C) report the data for both serotonin and tryptophan.Both tryptophan and ser otonin concentr ations r emained virtuall y unc hanged thr oughout the course of fermentation.The onl y significant incr ease in tryptophan pr oduction was detected in the OF treatment vessel at 24 h fermentation ( P = .004).Similarl y, the onl y significant incr eases in ser otonin pr oduction wer e detected on OF 0-24 h ( P = .016)and 10-24 h ( P = .015)fermentation ( Table S1 , Supporting Information ).
Finall y, no significant differ ences wer e detected in dopamine concentr ations thr oughout the entir e 48 h fermentation acr oss any of the substrates tested (Fig. 1 D; Table S1 , Supporting Inform ation ).

Organic acids
Figur e 2 (A)-(D) r eport concentr ations for acetate, pr opionate, butyr ate, and total or ganic acids thr oughout the course of fermentation.Lactate and succinate concentrations along with mean and specific donor data for all organic acids are reported in Tables S3  and S4 ( Supporting Information ).
Acetate was the most abundant organic acid detected r epr esenting between 53.9% and 69.2% of total organic acids produced at the end of fermentation.Acetate concentrations were highest in all treatments tested at 48 h and were all statistically significant compared with 0 h (Fig. 2 A; Table S3 , Supporting Information ).Increases in acetate production varied substantially between treatments with OF producing the largest acetogenic effect, averaging an increase of 91.94 ± 3.41 (SE) mmol/I abo ve baseline .Lowest increases in acetate production w ere recor ded on 2 FL avera ging an incr ease of 42.98 ± 3.94 (SE) mmol/I abo ve baseline .Increases in acetate in both OF and OF/2 FL treatments being statisticall y gr eater than sole 2 FL at 48 h (both P ≤ .001)( Table S3 , Supporting Information ).
Pr opionate pr oduction accounted for between 18.30% and 26.7% of total organic acids with all substrates inducing significant increases in propionate at 48 h compared with 0 h (Fig. 2 B).Most notable propiogenic substrate was OF, averaging an 33.82 ± 1.99 (SE) mmol/I incr ease abov e 0 h.The combination of OF/2 FL produced similar results inducing an average 31.21± 1.34 (SE) mmol/I incr ease abov e 0 h sampling ( Table S3 , Supporting Inform ation ).Lo w est increases in propionate were seen in the 2 FL treatment vessel at just 10.85 ± 1.66 (SE) mmol/I abo ve baseline .T he incr eases in pr opionate r ecor ded b y both OF and OF/2 FL combination being statistically different from sole 2 FL at 48 h (both P ≤ .001)(Fig. 2 B; Table S4 , Supporting Information ).
All substrates tested resulted in significant increases in butyr ate pr oduction, yet substantial differ ences in butyr ate pr oduction were seen between substrates.Most noticeable butyrogenic substr ates wer e OF and OF/2 FL with an av er a ge 22.22 ± 1.44 (SE) mmol/I and 18.65 ± 1.26 (SE) mmol/I increase above respective baseline samples .T he incr eases seen in butyr ate pr oduction in both OF and OF/2 FL tr eatments wer e statisticall y gr eater compared to 2 FL alone at 48 h (both P ≤ .001)(Fig. 2 C; Table S3 , Supporting Information ).
With respect to lactate, all substrates resulted in significant increases in lactate at 10 h fermentation mark ( Table S3 , Supporti ng Information ).Ther eafter, lactate concentr ations declined in all  S3 , Supporting Information ).
All substrates tested resulted in significant increases in succinate ( Table S3 , Supporting Information ).Yet, increases in succinate production varied substantially between substrates with OF and combination of OF/2 FL inducing significant increases in succinate production at 10 h (OF P ≤ .001)and (OF/2 FL P = .003),24 h (both P ≤ .001)and 48 h (both P ≤ .001),r espectiv el y ( Table S3 , Supporting Information ).In contrast, significant increases in succinate production in the 2 FL treatment vessel were only detected at 48 h ( P = .003).Furthermore, the increases detected in succinate concentrations at 48 h fermentation in both the OF and OF/2 FL tr eatment v essels wer e statisticall y differ ent fr om 2 FL alone (both P ≤ .001)( Table S3 , Supporting Information ).
Finall y, r egarding total or ganic acids, all substr ates r esulted in significant increases in total organic acids at the end of fermentation.OF induced the largest increases at 150.018 ± 5.38 (SE) mmol/I at 48 h and was statistically greater compared to 2 FL ( P ≤ .001),but not OF/2 FL.Smallest increases in total organic acids were recorded on 2 FL at 61.24 ± 4.02 (SE) mmol/I above 0 h (Fig. 2 D; Table S3 , Supporting Information ).

Bacterial enumer a tion
In order to determine shifts in bacterial populations, four 16S-rRNA fluorescence in situ hybridization probes were used to identify changes in numbers of total bacteria and three specifi-call y tar geted micr obial gr oups.Results of significant bacterial group counts during the batch culture fermentation are shown in Fig. 3 (A)-(D).Mean and donor specific data is reported in Tables S5  and S6 ( Supporting Information ).
Significant increases in total bacterial counts (Eub338-I, II, and III) wer e observ ed acr oss all substr ates with lar gest incr eases in microbial loads peaking at 24 h (Fig. 3 A).OF induced the largest increases in total bacteria counts averaging 0.72 log 10 ± 0.14 cells/ml above 0 h ( P ≤ .001).Smallest increases in total bacterial counts were induced on 2 FL and OF/2 FL combination at 0.52 log 10 ± 0.12 cells/ml and 0.55 log 10 ± 0.13 cells/ml above 0 h.Both of these were statistically significant from respective baseline samples (Fig 3 A; Table S5 , Supporting Information ).
Lar gest c hanges in micr obial numbers wer e r ecorded in Bifidobacterium (Bif164) counts with all three substrates tested recording significant increases across the course of fermentation.Yet, increases in Bif164 counts varied between the substrates tested with OF inducing the largest average increases in Bif164 counts at 1.49 log 10 ± 0.13 cells/ml (SE) ( P ≤ .001) at 24 h.Both 2 FL and the combination of OF/2 FL inducing similar increases of Bif164 counts at 1.22 log 10 ± 0.11 and 1.26 log 10 ± 0.11 (SE) cells/ml (Fig. 3 B; Table S5 , Supporting Information ).The changes seen in Bif164 counts correlating with those documented in acetate and GABA production throughout the course of fermentation.
All substrates resulted in significant increases in Bac303 counts.Significant differences were detected amongst treatments, with OF inducing lar gest incr eases in Bac303 at 24 h at 1.64 log 10  ± 0.23 cells/ml (SE) and remained significant until the end of fermentation (all P ≤ .001).In contrast, 2 FL induced the smallest increases in Bac303 counts at 0.99 log 10 ± 0.25 cells/ml (SE), both 8 h and 24 h samples recording significant differences compared to 0 h (Fig. 3 C; Table S5 , Supporting Information ).
Lastl y, significant differ ences in Lab158 counts wer e detected across all substrates tested (Fig. 3 D).Largest changes in Lab158 counts wer e r ecorded on OF with 0.8 log 10 ± 0.14 cells/ml (SE) increase at 8 h ( P ≤ .001),sole 2 FL recording similar av er a ge increases at 0.76 log 10 ± 0.17 cells/ml (SE) ( P ≤ .001).Smallest aver a ge incr eases in Lab158 counts were in the OF/2 FL treatment vessel at just 0.42 log 10 ± 0.21 cells/ml (SE) at 24 h fermentation (Fig. 3 D; Table S5 , Supporting Information ).Inter estingl y, within both the sole 2 FL and OF/2 FL combination fermentations sever al donors r ecorded higher av er a ge incr eases in Lab158 counts at 24 h fermentation, each of which were maintained until the end of fermentation ( Table S6 , Supporting Information ).

Discussion
In this in vitro batch culture investigation, we aimed to assess if the prebiotic OF and prebiotic candidate 2 FL, alone and in combination, could induce physiologically relevant changes in neuroactive metabolites and organic acid production.
Lar gest incr eases in neur otr ansmitters wer e seen in GABA pr oduction and were highest in OF and OF/2 FL fermentations peaking at 48 h, whereas smaller increases were detected on 2 FL peaking at 24 h, r espectiv el y, r esults being highly heterogenous amongst individual donors (Fig. 1 ; Tables S1 and S2 , Supporting Information ).Taking this into consideration it has been recently documented that rates of GABA synthesis can be dramatically influenced by the type of hexose and pentose sugars used during fermentation (Strandwitz et al. 2019, Cataldo et al. 2020 ).
Within the gut se v er al micr oor ganisms including Bifidobacterium , Bacteroides , and Lactobacillus have developed several mechanisms for producing GAB A. GAB A can be produced via the decarboxylation of l -glutamate or via the conversion of arginine to provide a protective mechanism on the low pH of the intestinal environment (Strandwitz et al. 2019, Otaru et al. 2021 ).We did not add arginine to our basal media thus the majority of GABA production likely occurred via the decarboxylation of l -glutamate.
Other mechanisms by which GABA production can occur include utilization of acetate and lactate as intermediates in the citric acid cycle (Rowlands et al. 2017 ).Consequentl y, giv en that ther e wer e significant incr eases in both acetate and lactate seen throughout the course of fermentation, specifically on OF and OF/2 FL, the additional increases seen in GABA in this study may be from utilization of both acetate and lactate via the gut microbiota.This coincides with results documented b y (F rost et al. 2014 ) who noted that 13 C-labelled acetate was able to cross the bloodbrain barrier regulating GABA production in the hypothalamus.This suggests that incr easing GABA pr oduction via increasing acetate pr oduction thr ough the use of the prebiotics ma y pro vide a means of increasing GABA production.This is of particular importance given GABA concentrations often correlate with levels of depression and several other mood state disorders (Brady et al. 2013, Al-Khafaji et al. 2020 ).
Small increases in both tryptophan and serotonin production wer e detected onl y on OF fermentations at 24 h (Fig. 1 B and C; Table S1 , Supporting Information ).While m uc h r emains unknown about the ability of the gut microbiota to produce tryptophan, dietary tryptophan can be utilized by the gut microbiota to produce se v er al deriv ativ e metabolites including indole and serotonin as well as entering the kynurenine pathway resulting in the production of kynurenic , quinolinic , or picolinic acid (Kaur et al. 2019, Gao et al. 2020 ).Yet, in this stud y, kyn urenic acid was under the limit of detection and the increases recorded in both serotonin and tryptophan production were relatively low compared to the increases seen in GABA production and likely do not reflect those seen in vivo .Taking this into consideration the conversion of tryptophan to serotonin occurs in the presence of enterochromaffin cells, ser otoner gic neur ons and necessary cofactors including pyridoxine-5-phosphate (Reigstad et al. 2015 ).Furthermore, while the pH used this study reflected the pH of the proximal colon, allowing for adequate exploration of GABA production by the micr obiota, it likel y did not provide optimal conditions for production of metabolites at the higher pH within the tr ansv erse and distal region of the colon.T herefore , it would be beneficial for future work to model proximal, transverse, and distal regions of the colon combined with the addition of all necessary precursors and cofactors in order to more adequately explore neurotransmitter production.
Concentrations of acetate , propionate , butyrate , lactate , and succinate all increased across the course of fermentation with OF and OF/2 FL in combination recording significantly greater incr eases compar ed to 2 FL (Fig. 2 ; Tables S3 and S4 , Supporting In formation ).
Metabolicall y deriv ed or ganic acids ar e speculated to play a pivotal role in context of the gut-brain axis being linked to se v er al health benefits.For example acetate, which was significantly increased in the OF and to a lesser extent OF/2 FL fermentations, has been linked to reductions in inflammation by regulating the expression of proinflammatory cytokines IL-6, TNF-α, and IL-1 β (Soliman et al. 2012, Underwood et al. 2015 ).Propionate is speculated to play a k e y role in regulating blood brain barrier permeability, along with pr otecting a gainst lipopol ysacc haride-mediated blood-brain barrier disruption (Braniste et al. 2014, Hoyles et al. 2018 ).Butyr ate, a gain significantl y enhanced in OF and OF/2 FL fermentations, has been shown to aid in regulating the expression of GABA r eceptors, enter oc hr omaffin cells, br ain-deriv ed neur otr ophic factor and glial-deriv ed neur otr ophic factor in mice and rats (Moris and Vega 2003, Intlekofer et al. 2013, Savignac et al. 2013, Reigstad et al. 2015 ) as well as being documented to decrease histone acetylation in piglets (Kien et al. 2008 ).
Or ganics acids, specificall y acetate has been linked to impr ov ements in satiety regulation via stimulation of leptin production in adipocytes (Zaibi et al. 2010 ), while succinate and propionate can act as precursors to gluconeogenesis (den Besten et al. 2013, Soty et al. 2015 ).
Significant increases in Bifidobacterium spp., Bacteroides / Prevotella , and Lactobacillus / Enterococcus were detected acr oss all substr ates with lar gest incr eases being documented in the OF fermentation (Fig. 3 A-D).In contrast, a high degree of heterogeneity was detected amongst donors in response to 2 FL, except when in the presence of OF where this heterogeneity a ppear ed to be mitigated.
The response of the gut microbiota, in particular Bifidobacterium spp., to OF is well-c har acterized with its effect being demonstrated in both in vitro (Wang andGibson 1993 , Pompei et al. 2008 ) and in vivo studies (Kolida et al. 2007, Vandeputte et al. 2017 ).The ability of the healthy adult gut microbiota to utilize 2 FL is extr emel y limited and our r esults ar e in line with those documented by (Yu et al. 2013, Suligoj et al. 2020, Ryan et al. 2021 ), indicating that a large responder/nonresponder status exists to 2 FL supplementation.
Taking this into consideration, the ability of the microbiota to utilize specific HMOs is documented to be highly species and e v en str ain specific (Lawson et al. 2020, Sakanaka et al. 2020, Jackson et al. 2022 ).This indicates that, if an individual is to utilize HMOs, they must possess the necessary microorganisms prior to supplementation.For example, within the 2 FL fermentations se v er al donors documented noticeable incr eases in Lab158 counts compared to OF fermentation (Fig. 3 D; Tables S5 and S6 , Supporting Information ).This is inter esting giv en that it has been pr e viousl y documented that Lactobacillus are not readily able to utilize intact HMOs to any real extent (Jackson et al. 2022 ).As ther e wer e also notable incr eases in Bifidobacterium within these individuals it is likely that increases in Lactobacillus occurred as r esult on scav enging on both lactose and fucose fr om the extr acellular degradation of HMOs by bifidobacteria (Zuniga et al. 2018 ).This further adds to the evidence that the mutualistic behaviour existing between micr oor ganisms found within the gut is a critical c har acter in helping to sha pe a div erse and flexible ecosystem (Jackson et al. 2022 ).
Our study is not, ho w e v er, without limitation.It should be acknowledged that the use of in vitro fermentation models to identify changes in microbial composition and resulting metabolites does not necessary ca ptur e c hanges seen in vivo .Furthermor e, it is important to note that while we did see significant increases in the production of several neurotransmitters evidence supporting the ability of gut-derived neurotransmitters to cross the bloodbrain barrier is not well-established (Strandwitz 2018 ) and findings should be inter pr eted as such.Ho w ever, in vitro models do allow for the testing of novel substrate combinations as prescreening tools for assessing potential changes in microbial composition and metabolite production prior to conducting human intervention trials and minimizing the heterogeneity seen in vivo .

Conclusion
Ov er all, the r esults of study impl y that OF and combinations of OF/2 FL were able to generate physiologically relevant increases in GABA and organic acids, all of which were maintained until the end of fermentation (all P ≤ .05),along with noticeable increases in Bifidobacterium , Bacteroides , and Lactobacillus counts.Whereas the ability of 2 FL to induce physiologically relevant increases in GABA and organic acid production along with substantiated changes in bacterial numbers appears to be highly donor specific, expect for when combined with OF, suggesting a strong r esponder/nonr esponder status exists.Additionally, OF was able to stimulate small increases in both serotonin and tryptophan, suggesting bacterial production of neuroactive metabolites may occur in the absence of colonic cells.Ho w e v er, these concentr ations were relatively low compared to GABA indicating that the gut microbiota is not the primary production pathway and that the presence of the necessary enteroendocrine cells is required.Ov er all, these r esults suggest that the pr ebiotic OF and combination of OF/2 FL should be taken forw ar d to a human intervention trial to determine their in vivo effects.

Figure 1 .
Figure 1.LC-MS analysis of neurotransmitter concentrations-(A) GABA, (B) serotonin, (C) tryptophan, and (D) dopamine in the supernatant of effluents collected from vessel 1-4 at 0, 10, 24, and 48 h representing mean ( n = 5) and standard error (SE) of the data with individual volunteer data points.Concentr ations r eported in (ng/ml).Results that ar e statisticall y significant within r espectiv e tr eatments ar e display ed b y specified P-values.Significant differences between treatments at 48 h are indicated by differing letters ( P ≤ .05).Abbr e viations: OF = oligofructose; 2 FL = 2 fucosyllactose.

Figure 2 .
Figure 2. GC-FID analysis of organic acid concentrations-(A) acetate, (B) propionate, (C) butyrate, and (D) total organic acids in the supernatant of effluents collected from vessel 1-4 at 0, 10, 24, and 48 h representing ( n = 5) of the data (all points).Concentration reported in (mmol/I) mean and SE.Results that are statistically significant within respective treatments are displayed by specified P-values.Significant differences between treatments at 48 h are indicated by differing letters ( P ≤ .05).Abbreviations: OF = oligofructose; 2 FL = 2 fucosyllactose.

Table 1 .
• C with a flow of 10 l/min.Injected sample v olume w as 4 μl.For the anal ysis of neur otr ansmitters, LC/MS Method P ac ka ge for Primary Metabolites (Shimadzu Cor por ation, K yoto , Japan) was used.The MRM transition for GABA was 104.10 > 87.05 m/z, tryptophan 205.10 > 188.15 m/z, serotonin 177.10 > 160.10 m/z, dopamine 154.10 > 91.05 m/z, kynurenic acid 190.10 > 144.10 m/z, nore pine phrine 170.10 > 152.15 m/z, and e pine phrine 184.10 > 166.10 m/z.Quantification of neur otr ansmitters was Name, sequence, and target group of oligonucleotide probe used in this study for FISH of bacterial enumeration.